Statistical Approach to Develop a Suitable Algorithm for Prediction of Apnea Using Heart Rate Variability Rather Than Other Conventional Methods

Abstract

Apnea is cessation of breathing with no movement of inspiratory muscles. There is no suitable way to detect the sleep apnea; doctors take the help of expensive and complicated whole night polysomnography or electroencephalogram (EEG) to diagnose sleep apnea. Research work is going on to detect obstructive sleep apnea (OSA) using different approaches (Mendonca et al. in A review of obstructive sleep apnea detection approach 2–8, 2018) [1]. In present research work, researchers are using heart rate variability as the main tool to predict different kinds of biological disorder in human beings (Lado et al. in Detecting sleep apnea by heart rate variability analysis: assessing the validity of databases and algorithms 1–2, 2009) [2]. Our present research work has focused to detect OSA using heart rate variability analysis (HRV) as a diagnostic tool. HRV is the measure of variation in the time duration between consecutive heartbeats in milliseconds. The main objective of our work is to improve accuracy and help doctors for proper diagnosis of apnea disorder. There are many critical evaluation processes of apnea, but here we are dealing with the simplest apnea detection process by HRV analysis using KUBIOS. In the beginning, we have started to collect the normal and abnormal (i.e., person with OSA) electrocardiogram (ECG) data from different subjects (Court-Fortune et al. in Eur Respir J 22:937–942, 2003) [3]. Our work proposes a high accuracy method to predict OSA based on HRV analysis. A significant difference in changing is observed in various parameters associated with heart rate variability of normal and apnea patient’s data, when different statistical parameters have been studied and compared. We have analyzed the heart rate variability using Kubios software (version 2.2). This work proposes few standard statistical parameters like root mean square of successive differences (RMSSD) or mean RR those will help to better prediction of OSA detection. Another parameters RR triangular index and TINN show a high difference in the range of results. Ultimately this work will instigate the research world that HRV can be used as a measuring tool for better prediction of OSA.